OPERATOR CONTROLLED DEVICE-TO-DEVICE COMMUNICATIONS IN LTE-ADVANCED NETWORKS

Transcription

1 ACCEPTED FROM OPEN CALL OPERATOR CONTROLLED DEVICE-TO-DEVICE COMMUNICATIONS IN LTE-ADVANCED NETWORKS LEI LEI AND ZHANGDUI ZHONG, BEIJING JIAOTONG UNIVERSITY CHUANG LIN, TSINGHUA UNIVERSITY XMIN (SHERMAN) SHEN, UNIVERSITY OF WATERLOO as gateway The authors study direct communications between user equipments in the LTE-advanced cellular networks. Different from traditional device-to-device communication technologies such as Bluetooth and WiFi-direct, the operator controls the communication process to provide better user experience and make profit accordingly. ABSTRACT This article studies direct communications between user equipments in the LTE-advanced cellular networks. Different from traditional device-to-device communication technologies such as Bluetooth and WiFi-direct, the operator controls the communication process to provide better user experience and make profit accordingly. The related usage cases and business models are analyzed. Some technical considerations are discussed, and a resource allocation and data transmission procedure is provided. INTRODUCTION Device-to-device (D2D) communications commonly refer to the technologies that enable devices to communicate directly without an infrastructure of access points or base stations, and the involvement of wireless operators. The term device here refers to the user who uses cell phones or other devices in Human-to- Human (H2H) communications as well as machine in Machine-to-Machine (M2M) communications without the involvement of human activities. The most widely known D2D technologies are Bluetooth and WiFi working at the 2.4GHz unlicensed band. Up to now, wireless operators don t include the D2D function in the universal cellular network standards, e.g., Global System for Mobile Communications (GSM), Universal Mobile Telecommunications System (UMTS) and 3rd Generation Partnership Project (3GPP) Long Term Evolution (LTE). This is largely because the D2D function was only envisioned as a tool to reduce the cost of local service provision, which is fractional according to the operators current market statistics. Recently, the wireless operators attitude towards the D2D function is changing because of several new trends in the mobile market. First, the context-aware applications are emerging in smart phones which are envisioned as an important value added service since an wireless operator can provide a plurality of services to a user according to its location information and its working status. For example, a user may be informed of a nearby restaurant, and the user can reserve a seat and get a coupon by making a call or sending a short message. Since most of the context-aware applications involve discovering and communicating with nearby devices, the D2D function can facilitate the discovery of neighboring devices and reduce the communication cost between these devices. Secondly, M2M applications are fast growing recently. Since the cellular equipments are getting smaller and cheaper, the wireless operators have great opportunities to connecting consumer electronic devices to their networks, e.g., washing machines and ovens. Since most consumer devices work around their owners, the cellular phone can be the hub for these devices and used as the gateway to the cellular networks. The D2D function enables the communications between consumer devices and cell phones. The above emerging services and applications are driving wireless operators to pursue the D2D function in their networks. However, the traditional D2D technologies are inadequate. First, there are more than 5 billion cellular users globally, who can only realize D2D function by WiFi or Bluetooth, which is not an integral part of the cellular networks and thus causes inconvenience customer usage experience. For example, both Bluetooth and WiFi require manual pairing between two devices. The distance of WiFidirect is claimed to be 656 inches, which means that dozens of devices within the range may be on the list. This process will make the user quite cumbersome compared to making a phone call. Second, the traditional D2D technologies are unable to meet the requirements of some users or applications due to several technical limitations. Since most of the traditional D2D technologies work on the crowded 2.4GHz unlicensed band, the interference is uncontrollable. In addition, traditional D2D technologies cannot provide security and Quality-of-Service (QoS) guarantee as the cellular networks. Last but not the least, the wireless operators cannot make profits using traditional D2D technologies /12/$ IEEE

2 since they work independently without the involvement of the operators. Unwilling to lose the emerging market that requires the D2D function, the wireless operators and vendors are exploring the possibilities of introducing the D2D function in the cellular networks. In [1], the concept of D2D communications as an underlay to an LTE-Advanced cellular network is introduced. A wireless technology called as the FlashLinQ that enables devices to directly sense their surroundings and directly communicate with each other is proposed in [2], which can be used in licensed band and as a complementary to the Wide Area Network (WAN). At the 3GPP meeting held in June 2011, a study item description on the radio aspects of device-to-device discovery and communication has been submitted by Qualcomm. Meanwhile, a study item description on LTEdirect is submitted to the 3GPP meeting held in August 2011, which proposes the study of the service requirement of direct over-the-air LTE device-to-device discovery and communication. Although interested in bringing the D2D function into cellular networks, the operators require to control the D2D services. Furthermore, the operator controlled D2D communications is facing a great dilemma in which if the users are charged for their D2D services, they may turn to traditional D2D technologies, which are free but with lower speed and less security. Therefore, the operators must answer the pay for what question before they can push forward the operator controlled D2D technology, which requires some analysis on the usage cases and business models. In this article, we first classify the operator controlled D2D communication technologies into two broad categories according to the level of operator control over D2D communications. The usage cases and business models are analyzed, followed by some technical considerations on the radio aspects of operator controlled D2D communications. Finally, the article is concluded. D2D CONTROLLED MODE The operator controlled D2D (OC-D2D) communications are defined as the technologies with which the devices communicate directly with each other under a cellular network or an operator control. The operator controls over normal user communication process which mainly lies in four aspects: access authentication, connection control, resource allocation, and lawful interception of communication information. The last aspect is very difficult to achieve for D2D communications, since information is directly exchanged between users bypassing the operator deployed base stations. According to the level of operator control over D2D communications, two categories of operator controlled D2D technologies can be classified. FULLY CONTROLLED D2D MODE The D2D link between two User Equipments (s) is an integral part of the cellular networks, just like the common cellular downlink or uplink connections. The cellular network has the full control over the D2D connection, including control plane functions, e.g., connection setup and maintenance, and data plane functions, e.g., resource allocations. The D2D connections share the cellular licensed band with the normal cellular connections. The network can either dynamically assign resources to each D2D connection in the same way as a normal cellular connection or semi-statically assign a dedicated resource pool to all D2D connections. The operator can charge the users for using D2D service based on how many minutes or how much bandwidth they use. LOOSELY CONTROLLED D2D MODE The operators perform the access authentication for the D2D enabled devices. Apart from this, these D2D devices can setup D2D connections and start D2D communication autonomously with little or no intervening from the operators. To avoid interference to the normal cellular users, the D2D communications can make use of either the unlicensed band with WiFi or Bluetooth for data transmission or a dedicated carrier on the licensed band. The operators can charge a certain amount of fee per month for providing the D2D service irrespective of the actual D2D data flow in the network. However, the operators must be able to disable the D2D service if the users do not pay for it. USAGE CASES AND BUSINESS MODELS The D2D usage cases can also be classified into two broad categories. The first category is referred to as the peer to peer case, in which the D2D devices are the source and destination of the exchanged data. The second category is the relay case, which means that one of the communicating D2D devices has to relay the exchanged information to the base station which further forwards the data to the destination device. PEER TO PEER Local Voice Service OC-D2D communications can be used to offload local voice traffic when two geographically proximate users want to talk on the phone, e.g., people in the same large meeting room want to discuss privately, or companions get lost in a supermarket, as shown in Fig. 1a. However, this usage case is rare according to the operators current market statistics. Local Data Service OC-D2D communications can also be used to provide local data service when two geographically proximate users or devices want to exchange data, as shown in Fig. 1b. Some scenarios of D2D communications are illustrated below. Content Sharing: Friends exchange photos or videos through their smart phones, or people attending a conference download materials from a local server. Multiplayer Gaming: The famous Japanese game Dragon Quest IX has a co-op mode consisting of up to four players using local wireless connections to play together. The three guests join the host system s world and can go anywhere that the host has explored. The operators can charge a certain amount of fee per month for providing the D2D service irrespective of the actual D2D data flow in the network. However, the operators must be able to disable the D2D service if the users do not pay for it. 97

3 DV Airport Supermarket Railway station Multiplayer gaming Internet Camera Local multicasting Machine to machine Meeting room Content sharing Context-aware application (a) (b) Sensors on home devices Computer as gateway Sensors on car Smart phone Sensors on customers Mobile internet devices (c) (d) Figure 1. D2D usage cases:a) local voice service; b) local data service; c) as gateway to sensor networks; and d) cooperative relay. Local Multicasting: The shops advertise the sale promotion information to the customers. Machine to Machine: A laptop connects to a printer, or a smart phone connects to a television for the photo or video display. Context-aware Application: It is a driving factor for the D2D technologies and is based on the people s desire to discover their surroundings and communicate with nearby devices (machines or people). One example is the Dragon Quest IX game, which has a tag mode allowing nearby game devices to discover each other and exchange messages automatically without the players awareness. Therefore, a player can take his game device to a mall or a coffee shop and find he has met a lot of interesting people when he comes back. Another example is location aware social networking, such as Foursquare, where users check-in at venues using a mobile website, text messaging or running a device-specific application and selecting from a list of venues that the application locates nearby. Each check-in awards the user points and sometimes badges. Therefore, contextaware applications may be based on any of the above four types of D2D communications, but they should be able to notice/interact when something nearby is interesting. Although it is appealing, making consumers adopt the OC-D2D technologies is tricky. The toughest problem is the competition with the traditional D2D technologies, which are currently dominating the market and allow the users to use local data service freely. Therefore, the OC- D2D technologies have to be attractive enough for consumers to switch and be willing to paying for this service. Some of the good reasons to attract users for operator controlled D2D services are listed below: Pay for identity: Use loosely controlled D2D technologies to link cellular phone number with WiFi or Bluetooth identity, which facilitates D2D connection setup and provides value-added service through identity management. For example, the users can start photo or video exchange through WiFi by dialing phone numbers instead of searching for the WiFi name; Pay for QoS and security: Use fully controlled D2D technologies for those services which require high QoS and/or security; Pay for context information: Operators have deep contextual information about end users and have an emerging opportunity to leverage context to make it pay off competitively. RELAY as the Gateway to Sensor Networks Most M2M devices are not directly cellular. In other words, M2M devices usually first connect to an M2M gateway using Wireless Personal Area Network (WPAN), e.g., zigbee, and the M2M 98

4 Usage Cases Peer to Peer Relay Local voice service local data service as gateway to sensor networks cooperative relay Benefits Enhance capacity Provide new services Enhance capacity Marketing Challenges Rare occasion Competition from traditional free D2D techniques Users concern on information security etc. Potential Business Models Attract users to pay for identity, QoS and security, context information, and management, etc. Charge the users based on how many minutes or how much bandwidth they use in fully controlled D2D communications; and charge a certain amount of fee per month irrespective of the actual D2D data flow in loosely controlled D2D communications Table 1. Usage cases for operator controlled D2D communications. gateway connects to a cellular network. For many consumer M2M devices, e.g., sensors on home devices, cars, or even the onbody health care devices, cell phones on the consumers are the most suitable M2M gateways. The communications between these sensors and s can use the OC-D2D technologies. Similar to the local data service usage case, the OC-D2D technologies also faces challenges from the traditional D2D technologies. The following are some possible answers to the pay for what question of the OC-D2D communications. Pay for management: One possible business model for the OC-D2D communications is given in [3], where non-cellular devices are included into the operator s subscriber database and automatically associated with the owner s cellular devices. An M2M profile is created for such device to store its relevant information, such as the owner and device specific access policies. Authentication/key management can be provided for the sensor devices that require security. An operator can also separately meter the data from different devices behind a phone/gateway. Pay for QoS and security: Use fully controlled D2D technologies for those applications that require high security and QoS, e.g., sensors for life care or security. Cooperative Relay In the wireless telecommunications systems that have a large number of subscribers, it is well known that one efficient communication method is to break a long path into a number of smaller hops so that the information is relayed between a number of terminals. The integration of cellular and ad-hoc networks to provide the relay capability has been well studied [4, 5]. 3GPP has even considered to apply this technique for the UMTS Time Division Duplex () under the concept of Opportunity Driven Multiple Access (ODMA) [6]. However, the relay faces a number of business model difficulties apart from the technical challenges. The biggest obstacle is the users concern on the information security, wireless radiation and excessive consumption of their battery power, all of which are due to opening up their mobile devices to other users. Therefore, 3GPP has finally decided to give up the ODMA standardization. These problems still exist today and a proper business model with enough incentives for users needs to be designed if the OC-D2D technologies is to be applied to this scenario. A further optimization for the relay is to apply the cooperative techniques [7], i.e., the D2D communications capability enables the users cooperation to achieve the transmit diversity, multi-antenna transmission and network coding, etc.. These techniques are still mostly under the academic research and not mature enough for the standardization and implementation in the near future. The benefits, marketing challenges and potential business models for different operator controlled D2D usage cases are summarized in Table 1. They should be taken into account in the design of various OC-D2D technologies. TECHNICAL CONSIDERATIONS IN RADIO ACCESS NETWORKS SPECTRUM FOR OPERATOR CONTROLLED D2D COMMUNICATIONS The categories of spectrum that the OC-D2D communications can operate on are listed as follows. Unlicensed band: The advantage is that operators do not need to sacrifice valuable licensed spectrum for providing D2D services. However, the uncontrolled interference condition makes this option rather unattractive to the users. One possible scenario of carrying D2D traffic on unlicensed band is that operators provide automatic device pairing, device authentication etc. using loosely controlled D2D technologies for the users. Frequency Division Duplex (FDD) licensed band: In order to support D2D function in FDD band, has to add Rx chain in the uplink spectrum or Tx chain in the downlink spectrum or both, which will increase cost and complexity. Therefore, for those operators who only have FDD spectrum, providing D2D service may be more difficult. licensed band: Since has both Rx 99

5 Power control and resource allocation of D2D connections can be either distributively determined by the s themselves or centrally performed by the base station, which is referred to as the Evolved Node B (enodeb) in LTE systems. downlink period uplink period FDD downlink FDD downlink Guard band Guard band carrier 1 carrier 1 Data transmission for D2D Data transmission for cellular Downlink control signaling for cellular and D2D s Uplink control signaling for cellular and D2D s Figure 2. Use of guard band between and FDD for D2D communications. and Tx chains for data transmission and reception in spectrum, no additional Rx/Tx chains are needed to support D2D communications, which makes spectrum more suitable for carrying D2D traffic. For operators with both and FDD spectrum, spectrum can be dedicated to D2D communications. For operators with only spectrum, D2D communications can share resources with normal cellular communications, or occupy one or several dedicated carriers. Guard band between FDD and : and FDD wireless systems that are deployed in the same geographical area need frequency separation referred to as guard band to prevent Radio Frequency (RF) interference with one another. Normally, guard band cannot be used for the sake of interference coexistence between and FDD. However, it is possible to transmit data on the guard band with properly designed techniques to increase resource utilization. For example, one approach is to apply half-duplex FDD in the guard band, where only uplink (or downlink) transmission is performed in the uplink (or downlink) period on the guard band between and FDD uplink (or downlink) carriers [8]. Using guard band for D2D communications can be more cost-efficient than using normal licensed band, which allows operators to provide low cost D2D services in competition with traditional D2D technologies. One possible method for transmitting the D2D service on the guard band between the and FDD uplink carriers is shown in Fig. 2, where the carrier aggregation technique is used to transmit the downlink control signaling from the base station to the D2D s on the carrier. POWER CONTROL, RESOURCE ALLOCATION AND INTERFERENCE MANAGEMENT Power control and resource allocation of D2D connections can be either distributively determined by the s themselves or centrally performed by the base station, which is referred to as the Evolved Node B (enodeb) in LTE systems. In the former case, dedicated resources carrier N carrier N Cellular Guard band Guard band D2D FDD uplink FDD uplink have to be allocated to all the D2D connections statically or semi-statically so that no interference should be caused to the cellular connections. In the latter case, the D2D connections may either use dedicated resources or share resources with cellular users, since the enodeb can make sure that the mutual interference between cellular and D2D connections are acceptable via scheduling and power control. An example of distributed resource allocation scheme is FlashLinQ, which is an Orthogonal Frequency Division Multiplexing (OFDM)-based synchronous MAC/PHY architecture for D2D communications. Unlike traditional D2D technologies, FlashLinQ is designed to work on licensed band, where interference is more controllable. The goal is to schedule a channel-state aware maximal independent set at any given time slot based on the current traffic and channel condition, and the scheduling algorithm leads to spatial throughput gains over an IEEE g system [9]. This resource allocation approach can be used in loosely controlled D2D communications on licensed band or fully controlled D2D communications where the base station semi-statically assigns a dedicated resource pool for all D2D users. Significant research has been done on the centralized resource allocation and power control algorithms considering mutual interference between D2D and cellular connections, where D2D communication is considered as an underlay to LTE-advanced networks [10 14]. The resource sharing mode selection problem, which decides whether the network shall assign D2D communication mode or not to a user pair and whether a pair of D2D users shall share resources with cellular users or use dedicated resources instead, is considered in [10, 11]. In [12, 13], resource allocation algorithms among the cellular and D2D links are studied. Reference [14] investigates power control algorithms for the D2D mode communications. Although the emphases of these works are different, the resource sharing mode selection, resource allocation and power control algorithms are usually considered jointly in order to achieve the optimal performance. 100

6 For the centralized resource allocation approach, the enodeb has full control over the resources allocated to each D2D connection and needs to inform the D2D s of the scheduled resources for data transmission via L1/L2 control signaling, e.g., Physical Downlink Control Channel (PDCCH). However, this problem has not been adequately addressed. In this article, we provide a resource allocation and data transmission procedure, as shown in Fig. 4. This procedure gives an example of how a centralized resource allocation approach could be implemented in an LTE-advanced system, identifying the possible control and data channels between the enodeb and the D2D s for control signaling and data service transmission. Assume 1 and 2 have established an D2D connection and 1 has data wait to be transmitted to 2. The enodeb is responsible for resource allocation. First, 1 notifies the enodeb that it has data to be transmitted to 2. According to the LTE protocol, 1 can send a buffer status report (BSR) [15] to the enodeb through the Physical Uplink Shared Channel (PUSCH) [16] for this purpose. If no uplink resources are available for the BSR transmission, 1 can send a one bit scheduling request (SR) [15] signaling through the Physical Uplink Control Channel (PUCCH) [16]. Once the enodeb receives the SR from 1, it will allocate a small amount of uplink resources for the BSR transmission. After the enodeb receives the notification (e.g., BSR) from 1, it will allocate resources for the data transmission between 1 and 2. The specific resource allocation algorithms will not be discussed here. In an LTE system, the enodeb usually considers the channel status when performing resource allocation. For the D2D communications, the enodeb can obtain the channel status of D2D links between 1 and 2 by the periodic or aperiodic channel quality indication (CQI) [17] reports from 1 and 2 through the PUCCH. It is assumed that 1/2 can perform the CQI estimation from the received Sounding Reference Signal (SRS) [16] transmitted by its D2D peer. Once the enodeb determines the allocated resources, it notifies the result to both 1 and 2 through the PDCCH. In an LTE system, a performs blind decoding using its identity (i.e., Cell Radio Network Temporary Identifier (C-RNTI) [18]) to locate the specific PDCCH for it [17]. Therefore, in order to simultaneously notify 1 and 2 the resource allocation result, two possible methods are described below. Method 1: The enodeb sends two independent PDCCHs to 1 and 2 with their own C-RNTIs. In order for 1 and 2 to know whether it should transmit or receive data on the allocated resources, the PDCCHs that the enodeb sends to 1 and 2 can be in different Downlink Control Information (DCI) formats [19]. In this example, the enodeb sends Uplink (UL) grant to the sending 1 and Downlink (DL) grant to the receiving 2. Method 2: The enodeb sends only one PDCCH to 1 and 2 with the C-RNTI of the sending (1 in this example). Therefore, 2 needs to know the C-RNTI of 1 in order to decode this PDCCH. This can be obtained during the D2D connection establishment phase. Compared with Method 1, this approach can reduce the signaling overhead but increase the blind decoding attempts. After 1 receives the PDCCH from the enodeb, it will transmit data to 2 on the allocated resources. In an LTE system, downlink and uplink data transmissions are carried on the Physical Downlink Shared Channel (PDSCH) [16] and the PUSCH, respectively. In the D2D communications, however, there is no differentiation between downlink and uplink since the two communicating devices are both s. Therefore, it seems that both the PDSCH and the PUSCH can be used to carry the D2D traffic. However, there are several issues if the D2D links are considered as downlink. In an LTE system, a needs to estimate the downlink channel by detecting the Cell-specific Reference Signals (CRS) [16] from the enodeb to carry out downlink coherent demodulation. Since a D2D cannot transmit the CRS as an enodeb does, or mutual interference between the enodeb and the transmitting will arise and the signal demodulation may not be performed correctly. In addition, the PDCCH has to be transmitted one or several subframes prior to the PDSCH in the D2D case, which will cause the cross-subframe scheduling problem. On the other hand, it is more straightforward to consider the D2D links as uplink. In the current LTE systems, the uplink reference signals are -specific, so there is no interference problem for the reference signals from a D2D and a cellular. In addition, the enodeb notifies the its scheduled uplink resources one or several subframes prior to the PUSCH transmission by the, the timing relationship between the PDCCH and PUSCH are suitable for the D2D case and there shall be no cross-subframe scheduling problem. After 2 receives the PDCCH from the enodeb, it will receive data on the allocated resource from 1, which are transmitted through the PUSCH as discussed above. It is assumed that 2 has the Single-Carrier Frequency Division Multiple Access (SC-FDMA) baseband reception ability and uplink Demodulation Reference Signal detection ability. 2 then provides an ACK/NACK feedback to 1 according to whether the data is correctly received or not. In an LTE system, the enodeb transmits an ACK/NACK to a for its PUSCH transmission by the Physical Hybrid ARQ Indicator Channel (PHICH) [16], which is mapped to the first three OFDM symbols. However, since 2 may simultaneously transmit data to 1 by the PUSCH, which cannot be multiplexed in the same subframe with the PHICH, it is proposed to use the PUCCH for the ACK/NACK transmission instead. In an LTE system, the PUCCH is used to carry ACK/NACK for downlink data transmission. However, the physical resources that the PUCCH is mapped to are related to the PDCCH that schedules the corresponding PDSCH. Unlike cellular connections, where the PDCCH Significant research has been done on the centralized resource allocation and power control algorithms considering mutual interference between D2D and cellular connections, where D2D communication is considered as an underlay to LTE-advanced networks. 101

7 enb PUCCH (SR) PDCCH (UL grant for BSR) PUCCH (BSR) PDCCH (UL grant) 1 PDCCH (UL grant) PUSCH (initial transmission) PUCCH (ACK/NACK) PUSCH (retransmission) Figure 3. Resource allocation and data transmission procedure for D2D communications. 2 and the related PDSCH are transmitted in the same subframe, the PDCCH has to be transmitted one or several subframes prior to the actual data transmission for the D2D connections. Therefore, the PUCCH resource for D2D ACK/NACK may collide with a cellular ACK/NACK, since their corresponding PDC- CHs are transmitted in different subframes. In order to solve this problem, two possible methods are provided. Method 1: Reserve specific resources for the D2D PUCCH. Method 2: Use the cross-carrier scheduling capability in the carrier aggregation technique, where a dedicated carrier is used to carry the D2D traffic and the related PDCCHs are carried on the other carriers for cellular traffic. In this way, the D2D ACK/NACK will be transmitted on a different carrier from the cellular ACK/NACK, which will avoid collision between them. One example is to use the guard band of and FDD for D2D traffic. Finally, when 1 receives the ACK/NACK from 2, it will decide whether to perform data retransmission or not. Since the LTE system uses synchronous HARQ in the uplink, 1 and 2 both know on which subframes to send and receive the retransmitted data. In addition, if non-adaptive HARQ is adopted, 1 can retransmit the data on the same resource blocks as the initial transmission, so that the enodeb does not need to send the PDCCH to 1 and 2 again. However, the enodeb has to listen to the ACK/NACK from 2 to determine whether it can schedule new data on these resources. The control and data channels for the above D2D communication procedure is shown in Fig. 3. Since this article only discusses about the radio aspects of D2D communications, the resource allocation issue in core networks is out of our scope. However, it should be noted that no core network resources are needed to carry D2D traffic. PEER DISCOVERY, PAGING AND CONNECTION ESTABLISHMENT Before the resource allocation and data transmission phase, two devices need to find each other, i.e., peer discovery and D2D connection setup. The peer discovery phase is relatively independent of the D2D communication phase. Existing work can be classified into centralized and distributed approaches. Centralized approach: A certain entity in the cellular network, e.g., Packet Data Network (PDN) gateway or Mobility Management Entity (MME), detects that it may be better for two communicating s to set up a D2D connection. This entity then informs the enodeb to request measurements from the to check if the D2D communications offers higher throughput. If so, the enodeb decides that the two s can communicate in D2D mode [1]. Distributed approach: The broadcasts identity periodically so that other s may be aware of its existence and decides whether it shall start a D2D communication with it. This approach does not need the involvement of the base station [2]. The distributed peer discovery approach is more flexible and scalable than the centralized one. However, the operator cannot forbid illegal users to announce or listen information to/from the D2D peers using the operators licensed band. For fully controlled and loosely controlled D2D communications, different paging and connection establishment methods may be used: Fully controlled D2D communications: The paging and connection establishment procedure is mostly the same with normal LTE procedure [20]. However, since the D2D s shall exchange data directly over the air after the connections are established, it may be necessary to inform one D2D of the configured information of its peer regarding data transmission, e.g., C-RNTI, sounding reference signal configuration, and ciphering key etc. Loosely controlled D2D communications on licensed band: When communication between two s is desired, contact can be initiated via a form of direct D2D paging to create a D2D connection without the intervention of the base station [2]. However, the operator should be able to control whether the D2D connection is allowed to be setup or not. Loosely controlled D2D communications on unlicensed band: The cellular network perform authentication when two s want to start D2D communication. After that, the data transmission between these s takes place on unlicensed band with traditional D2D technologies. CONCLUSION In this article, we have studied the potential usage cases and technical design considerations in the operator controlled device-to-device communications. The potential usage cases have been analyzed and classified into four categories. Each usage case has its own marketing challenges and the design of the related techniques should take these factors into consideration. Furthermore, some technical considerations on the radio aspects of the operator controlled 102

8 D2D communications are discussed, including the usable spectrum, resource allocation and connection establishment, etc. Specifically, a signaling procedure of the resource allocation and data transmission of operator controlled D2D communications has been provided. The operator controlled D2D communications should enable the operators to control their networks in order to provide better user experience and make profit accordingly. At the same time, they should be flexible and low-cost to compete with traditional free D2D communications. The operators still face several challenges in providing such a D2D solution that can address the above two contradicting objectives simultaneously. ACKNOWLEDGEMENT This work was supported by the National 973 Projects (No. 2010CB328105, No. 2009CB320504), the Fundamental Research Funds for the Central Universities (No. 2012JBM003), and the Key Project of the National Natural Science Foundation of China (No ). REFERENCES [1] K. Doppler et al., Device-to-Device Communication as an Underlay to LTE-Advanced Networks, IEEE Commun. Mag., vol. 47, no. 12, Dec. 2009, pp [2] M. Scott Corson et al., Toward Proximity-Aware Internetworking, IEEE Wireless Mag., vol. 17, no. 6, Dec. 2010, pp [3] H. Viswanathan, Expanding the Role of the Mobile Network Operator in M2M, 1st ETSI TC M2M Wksp., Sophia Antipolis, France, Oct [4] H. Fitzek, M. Katz, and Q. Zhang, Cellular Controlled Short-Range Communication for Cooperative P2P Networking, Wireless Pers. Commun., vol. 48, no. 1, Jan. 2008, pp [5] H. Wu et al., An Integrated Cellular and Ad hoc Relaying System: icar, IEEE JSAC, vol. 19, no. 10, Oct. 2001, pp [6] 3GPP TR v1.0.0, Opportunity Driven Multiple Access, Dec [7] F. Fitzek and M. Katz, Cooperation in Wireless Networks: Principles and Applications, Springer Netherlands, [8] E. F. Gormley and C. A. Pralle, Utilizing Guard Band between FDD and Wireless Systems, US Patent Application , June [9] X. Wu et al., FlashLinQ: A Synchronous Distributed Scheduler for Peer-to-Peer Ad Hoc Networks, 48th Annual Allerton Conf., Illinois, USA, Sept [10] C.-H. Yu et al., Resource Sharing Optimization for Device-to-Device Communication Underlaying Cellular Networks, IEEE Trans. Wireless Commun., vol. 10, no. 8, Aug. 2011, pp [11] S. Hakola et al., Device-to-Device (D2D) Communication in Cellular Network Performance Analysis of Optimum and Practical Communication Mode Selection, Proc. IEEE Wireless Commun. and Net. Conf., [12] M. Zulhasnine, C. Huang, and A. Srinivasan, Efficient Resource Allocation for Device-to-Device Communication Underlaying LTE Network, Proc. IEEE Wireless Commun. and Networking Conf., [13] P. Hanis et al., Interference-Aware Resource Allocation for Device-to-Device Radio Underlaying Cellular Networks, IEEE 69th Vehic. Tech. Conf. Spring, [14] H. Xing, and S. Hakola, The Investigation of Power Control Schemes for A Device-to-Device Communication Integrated into OFDMA Cellular System, IEEE 21st Int l. Symp. Personal Indoor and Mobile Radio Commun., [15] 3GPP TS v10.3.0, Evolved Universal Terrestrial Radio Access (E-UTRA); Medium Access Control (MAC) protocol specification, Sept [16] 3GPP TS v10.3.0, Evolved Universal Terrestrial Radio Access (E-UTRA); Physical channels and modulation (Release 10), Sept PDCCH (UL grant) PUCCH (SR, CQI) PUSCH (BSR) PUSCH (data) PUCCH (ACK/NACK) enodb Figure 4. Data and control channels for D2D communications. [17] 3GPP TS v10.3.0, Evolved Universal Terrestrial Radio Access (E-UTRA); Physical layer procedures (Release 10), Sept [18] 3GPP TS v10.5.0, Evolved Universal Terrestrial Radio Access (E-UTRA) and Evolved Universal Terrestrial Radio Access Network (E-UTRAN); Overall description; Stage 2 (Release 10), Sept [19] 3GPP TS v10.3.0, Evolved Universal Terrestrial Radio Access (E-UTRA); Multiplexing and channel coding (Release 10), Sept [20] 3GPP TS v10.3.0, Evolved Universal Terrestrial Radio Access (E-UTRA); Radio Resource Control (RRC); Protocol specification (Release 10), Sept BIOGRAPHIES LEI LEI ([email protected]) received a B.S. degree in 2001 and a Ph.D. degree in 2006, respectively, from Beijing University of Posts & Telecommunications, China, both in telecommunications engineering. From July 2006 to March 2008, she was a postdoctoral fellow at Computer Science Department, Tsinghua University, Beijing, China. She worked for the Wireless Communications Department, China Mobile Research Institute from April 2008 to August She has been an Associate Professor with the State Key Laboratory of Rail Traffic Control and Safety and the School of Electronic and Information Engineering, Beijing Jiaotong University, since Sept Her current research interests include performance evaluation, quality-of-service and radio resource management in wireless communication networks. CHUANG LIN [SM] ([email protected]) is a professor of the Department of Computer Science and Technology, Tsinghua University, Beijing, China. He is a Honorary Visiting Professor, University of Bradford, UK. He received the Ph.D. degree in Computer Science from the Tsinghua University in His current research interests include computer networks, performance evaluation, network security analysis, and Petri net theory and its applications. He has published more than 400 papers in research journals and IEEE conference proceedings in these areas and has published four books. He is a senior member of the IEEE. He serves as the Technical Program Vice Chair, the 10th IEEE Workshop on Future Trends of Distributed Computing Systems (FTDCS 2004); the General Chair, ACM SIGCOMM Asia workshop 2005 and the 2010 IEEE International Workshop on Quality of Service (IWQoS PUCCH (CQI) PDCCH (UL grant) 103

9 2010); the Associate Editor, IEEE Transactions on Vehicular Technology; and the Area Editor, Journal of Computer Networks. XMIN (SHERMAN) SHEN [M 97, SM 02, F 09] ([email protected]) received the B.Sc.(1982) degree from Dalian Maritime University (China) and the M.Sc. (1987) and Ph.D. degrees (1990) from Rutgers University, New Jersey (USA), all in electrical engineering. He is a Professor and University Research Chair, Department of Electrical and Computer Engineering, University of Waterloo, Canada. He was the Associate Chair for Graduate Studies from 2004 to His research focuses on resource management in interconnected wireless/wired networks, wireless network security, wireless body area networks, vehicular ad hoc and sensor networks. He is a coauthor/editor of six books, and has published more than 600 papers and book chapters in wireless communications and networks, control and filtering. He served as the Technical Program Committee Chair for IEEE VTC 10 Fall, the Symposia Chair for IEEE ICC 10, the Tutorial Chair for IEEE VTC 11 Spring and IEEE ICC 08, the Technical Program Committee Chair for IEEE Globecom 07, the General Co- Chair for Chinacom 07 and QShine 06, the Chair for IEEE Communications Society Technical Committee on Wireless Communications, and P2P Communications and Networking. He also serves/served as the Editor-in-Chief for IEEE Network, Peer-to-Peer Networking and Application, and IET Communications; a Founding Area Editor for IEEE Transactions on Wireless Communications; an Associate Editor for IEEE Transactions on Vehicular Technology, Computer Networks, and ACM/Wireless Networks, etc.; and the Guest Editor for IEEE JSAC, IEEE Wireless Communications, IEEE Communications Magazine, and ACM Mobile Networks and Applications, etc. He received the Excellent Graduate Supervision Award in 2006, and the Outstanding Performance Award in 2004, 2007 and 2010 from the University of Waterloo, the Premier s Research Excellence Award (PREA) in 2003 from the Province of Ontario, Canada, and the Distinguished Performance Award in 2002 and 2007 from the Faculty of Engineering, University of Waterloo. He is a registered Professional Engineer of Ontario, Canada, an Engineering Institute of Canada Fellow, and a Distinguished Lecturer of IEEE Vehicular Technology Society and Communications Society. ZHANGDUI ZHONG ([email protected]) received the B.Eng. and M.Sc. degrees from Northern Jiaotong University (currently Beijing Jiaotong University), Beijing, China, in 1983 and 1988, respectively. He has been a Professor with the School of Electronic and Information Engineering, Beijing Jiaotong University, since He has authored seven books and over 150 technical papers in the field of wireless communication. His current research interests include wireless communication theory for railway systems, wireless ad hoc networks, channel modeling, radio resource management, intelligent transportation systems, and GSM- R systems. 104